University of Pennsylvania develops AI interface for epilepsy monitoring

Researchers at the University of Pennsylvania have developed a conversational artificial intelligence platform that allows epilepsy patients to communicate in real time with their brain monitoring devices. The study, published in Nature Medicine, involved 13 adults in a clinical trial at the University of Pennsylvania Hospital.

By utilizing a smartphone application connected to a secure cloud environment, patients can query their physiological data, request visual trends, or report symptoms. This bidirectional communication allows patients to provide immediate feedback to the electroencephalography (EEG) devices, marking detections as true, false, or uncertain. This feedback loop reduced false seizure alarms by approximately 80 percent.

The system also demonstrated high responsiveness, answering over 92 percent of patient queries accurately. Experts noted that while these devices are primarily designed for clinicians, providing patients with access to their own data offers significant potential for more adaptive disease management.